Merge remote-tracking branch 'origin/textequiv-index'

ocrd_page_to_alto/cli.py

@@ -16,7 +16,9 @@
 @click.option('--dummy-textline/--no-dummy-textline', default=True, help='Whether to create a TextLine for regions that have TextEquiv/Unicode but no TextLine')
 @click.option('--dummy-word/--no-dummy-word', default=True, help='Whether to create a Word for TextLine that have TextEquiv/Unicode but no Word')
 @click.option('--textequiv-index', default=0, help='If multiple textequiv, use the n-th TextEquiv by @index')
-@click.option('--textequiv-fallback-strategy', default='last', type=click.Choice(['raise', 'first', 'last']), help="What to do if nth textequiv isn't available. 'raise' will lead to a runtime error, 'first' will use the first TextEquiv, 'last' will use the last TextEquiv on the element")
+@click.option('--textequiv-fallback-strategy', default='first', type=click.Choice(['raise', 'first', 'last']), 
+              help="What to do if selected TextEquiv @index is not available: 'raise' will lead to a runtime error, "
+              "'first' will use the first TextEquiv, 'last' will use the last TextEquiv on the element")
 @click.option('--region-order', default='document', help="Order in which to iterate over the regions", type=click.Choice(['document', 'reading-order', 'reading-order-only']))
 @click.option('--textline-order', default='document', help="Order in which to iterate over the textlines", type=click.Choice(['document', 'index', 'textline-order']))
 @click.option('-O', '--output-file', default='-', help='Output filename (or "-" for standard output, the default)',

ocrd_page_to_alto/utils.py

@@ -38,16 +38,16 @@ def get_nth_textequiv(reg_page, textequiv_index, textequiv_fallback_strategy):
         if textequiv_fallback_strategy == 'raise':
             raise ValueError("PAGE element '%s' has no TextEquivs and fallback strategy is to raise" % reg_page.id)
         return ''
-    if len(textequivs) < textequiv_index + 1:
-        if textequiv_fallback_strategy == 'raise':
-            raise ValueError("PAGE element '%s' has only %d TextEquiv elements so cannot choose the %s%s and fallback strategy is to raise" % (
-                reg_page.id, len(textequivs), textequiv_index + 1, 'st' if textequiv_index == 0 else 'nd'))
-        elif textequiv_fallback_strategy == 'first':
-            return textequivs[0].Unicode
-        else:
-            return textequivs[-1].Unicode
+    for textequiv in textequivs:
+        if textequiv.get_index() == textequiv_index:
+            return textequiv.Unicode
+    if textequiv_fallback_strategy == 'raise':
+        raise ValueError("PAGE element '%s' has no TextEquiv index %d" % (
+            reg_page.id, textequiv_index))
+    elif textequiv_fallback_strategy == 'first':
+        return textequivs[0].Unicode
     else:
-        return textequivs[textequiv_index].Unicode
+        return textequivs[-1].Unicode
 
 def contains(el, bbox):
     minx1, miny1, maxx1, maxy1 = bbox